How a Consumer-Goods Leadership Team Cut Accident Ratio 50% in 6 Months by Changing Supervisor Tolerance

The PepsiCo South America case is useful because the headline number was simple. Accident ratio fell 50% in 6 months. The real change was less visible. Supervisors stopped treating small deviations as harmless, and that changed what the field expected from leadership.

This is not a story about perfect behavior. It is a story about a tighter boundary. Across 25+ years in multinational EHS and more than 250 cultural transformation projects, Andreza Araújo has seen the same pattern repeat: when supervisors tolerate the shortcut, the shortcut becomes normal. When supervisors stop tolerating it, the field changes faster than a campaign can move it. As Andreza Araújo argues in Safety Culture: From Theory to Practice, culture is not an announcement. It is the sum of repeated decisions under pressure.

The case sits in safety leadership because the result did not come from a slogan, a poster, or a one-off push. It came from leadership behavior that made the standard more immediate at the point of work. That is also why James Reason is relevant here. Latent failures rarely begin with a dramatic event. They begin when the system teaches people that the small deviation is acceptable long before it becomes visible harm.

Initial scenario

The starting point was familiar in many operations. Safety was declared as a priority, yet the daily rhythm still allowed small deviations to pass when the shift was busy. A permit could be read too quickly. A handover could leave a weak signal unchallenged. A supervisor could decide that production pressure justified one more exception. None of those choices looked serious in the moment, but together they taught the team what the real rule was.

That is why many operations miss the problem. They look at the written rule and assume the work follows it. The field usually follows the rule that the supervisor enforces, especially when the job is noisy, rushed, or uncomfortable. In Andreza Araújo's book Make The Difference: Be a Leader in Health & Safety, that boundary is treated as a core leadership duty, not as a side topic. If the leader does not defend the standard in the moment it is challenged, the standard remains theoretical.

The important shift in this case was not a new technical control. It was a new answer to a hard question. What is the organization willing to let pass when the shift is under pressure? Once that question was answered differently, the field began to notice.

Decision

The leadership decision was to move the standard closer to the shift. Instead of treating safe behavior as a campaign subject, the organization made it part of the supervisor role. The supervisor was no longer just a messenger for production. The supervisor became the first owner of the boundary between acceptable work and a shortcut that could not be normalized.

That sounds obvious, but many sites still split responsibility from authority. EHS defines the rule, operations carries the schedule, and the supervisor sees the shortcut in real time. If the supervisor is not expected to intervene, the rule stays in the slide deck while the work keeps drifting. The PepsiCo South America case shows that the real decision was to stop tolerating that split.

In more than 250 cultural transformation projects, Andreza Araújo has seen that leaders often overestimate the power of communication and underestimate the power of permission. A team does not only learn from what is said. It learns from what is allowed. That is why the decision was structural. It changed who had to answer when the field stepped toward an exception.

This also fits Patrick Hudson's maturity logic in a practical sense. A mature system does not merely know the rule. It reacts consistently when the rule is about to be broken. That consistency is what creates trust in the standard.

Execution

Execution depended on rhythm, not drama. Supervisors had to walk the work, see where shortcuts usually appeared, and intervene before the exception became routine. If a permit was rushed, it was paused. If a handover was thin, it was reopened. If task design pushed people toward a deviation, the condition was escalated instead of blaming the worker alone.

That style of execution matters because teams learn from the first correction. If the first correction is soft, late, or inconsistent, the field concludes that the old tolerance still exists. If the correction is clear and immediate, the team adjusts quickly because the risk is now visible in the supervisor's response. That is the difference between a message and a control.

The same discipline appears in Leadership Cadence: How 180 Days Cut Accidents, where the operating rhythm matters more than the speech. It also connects with New Plant Manager in 90 Days: Safety Leadership Plan, because a new leader who does not define tolerance early usually inherits someone else's boundary.

Execution worked because it was visible at the point of work. People could see that the same deviation no longer received the same pass. That visibility is more important than a polished campaign, because the field changes when the local decision changes, not when the communication team finishes a message.

Measured result

The measured result in the PepsiCo South America case was a 50% reduction in accident ratio in 6 months, according to Andreza Araújo's field record from that transformation. The number matters, but the deeper result is more useful. The organization made unsafe tolerance harder to reproduce, and that changed the trajectory of harm faster than a campaign would have done on its own.

Before and after makes the shift easier to read.

That table is important because it shows the control system, not just the outcome. James Reason helps explain why. Latent failure is not only about the final event. It is about the conditions that make a weak choice easy to repeat. This case changed those conditions by making tolerance visible and less permissive.

The result also fits Andreza Araújo's broader record across 30+ countries, because the same pattern shows up in different languages and different operating models. The strongest change usually comes when supervisors, managers, and EHS teams stop sending mixed signals about what is allowed.

Generalizable lessons

Lesson 1. Safe behavior is governed locally. Workers learn the real rule from the supervisor who sees the work in motion, not from the poster in the corridor. If the supervisor tolerates a deviation once, the team will test whether that tolerance is permanent. That is why the line of supervision matters more than the line on the slide.

Lesson 2. Behavior follows certainty. When people know the standard will be defended, they adjust faster than when they think the rule is negotiable. A leader who acts the same way every time creates a predictable boundary, and predictability lowers the cost of doing the right thing.

Lesson 3. Metrics improve after behavior changes, not before. Leaders often try to reverse that order, which is why dashboards can get cleaner while the work remains fragile. The PepsiCo South America case shows the opposite sequence. The supervisor boundary changed first, then the ratio moved.

Lesson 4. Pressure exposes the real culture. A site may look disciplined when the schedule is calm, yet the true norm appears when time is short and work is messy. If tolerance rises with pressure, the organization has not built a stable standard. It has built a flexible excuse.

Lesson 5. Leadership is a field action. The organization did not need more abstract language about values. It needed fewer exceptions that could pass as normal. Andreza Araújo's book A Ilusão da Conformidade, glossed in English as The Illusion of Compliance, is relevant here because many sites look orderly until someone checks what really happens under load.

What to apply in your operation

If you lead a plant, a region, or a contractor-heavy site, start with one question this week. What behavior do supervisors still tolerate because it looks small? Then name the exact point where that behavior enters the work, and make the supervisor the first owner of the response.

From there, check three things in the field. First, whether the standard is visible at the point of work. Second, whether the supervisor can stop the task without waiting for permission from a meeting. Third, whether follow-up proves that the deviation was removed instead of merely discussed. Those checks are simple, but they force the system to show whether tolerance still hides in the routine.

This is where the case becomes practical for a new plant manager or EHS leader. If you want the boundary to move, you do not start with a campaign calendar. You start with the supervisor, the moment of intervention, and the speed of follow-up. That is the same logic behind Safety Culture Diagnosis: 250-Company Case, because diagnosis only matters if it changes the decisions made in the shift.

One useful field test is to watch the first correction of the day. If it is vague, delayed, or negotiated, the site still tolerates drift. If it is clear and immediate, the field has learned that the standard is real. That is the operating difference between declared safety and operated safety.

Conclusion

This case shows that a 50% reduction in accident ratio in 6 months can begin with something less visible than technology or budget. It can begin when supervisors stop tolerating the shortcuts that used to define normal work, because behavior changes when the field learns that the standard will be defended every day.

If your operation needs help turning tolerated deviation into visible leadership discipline, start at Andreza Araujo. The goal is not to sound strict. The goal is to make safe behavior the easier option in the shift that is happening right now.